The DHS Program User Forum      
Discussions regarding The DHS Program data and results
Home » Topics » Nutrition and Anthropometry » Applying weights in a multi-country analysis
Applying weights in a multi-country analysis [message #10292] Wed, 20 July 2016 08:35 Go to next message
ttuti is currently offline  ttuti
Messages: 4
Registered: March 2016
Member
Hello,

I am conducting research into child malnutrition and I am using multi-level analysis. The sample is from Kenya 2014,Rwanda 2014,Ghana 2014, Namibia 2013 and Senegal 2014. However, I am unsure if my application of weights is appropriately used in my analysis. For the pooled analysis, I use the following formula:

Weights = (V005/100000) * (Country sample size/Pooled Sample Size)

The purpose of using weights in my analysis is to account for the unequal probability of household selection. Is the use of my weights appropriate? If not, how do I remedy it?

Best wishes,
Tuti.
Re: Applying weights in a multi-country analysis [message #10379 is a reply to message #10292] Thu, 21 July 2016 19:50 Go to previous messageGo to next message
Reduced-For(u)m
Messages: 269
Registered: March 2013
Senior Member

I THINK that what you are trying to do is to make the weights sum up to 1 for each country (implicitly preserving probability of selection within a survey, and making each survey have the same total weight). If that is right, you want to divide each weight by the sum of WEIGHTS for each survey (not the sample size). Does that help?
Re: Applying weights in a multi-country analysis [message #10381 is a reply to message #10379] Fri, 22 July 2016 05:06 Go to previous message
ttuti is currently offline  ttuti
Messages: 4
Registered: March 2016
Member
Yes, that does help alot. Thank you for the help.
Previous Topic: Missing values for child weight and height
Next Topic: Exclusive breastfeeding
Goto Forum:
  


Current Time: Fri Oct 20 04:48:11 Eastern Daylight Time 2017